The specific aims of the project are: 1. to describe the changes in provision of rehabilitation services to newly admitted nursing home patients following the implementation of the Nursing Home (NH) Prospective Payment System (PPS) 2. to detail changes in rehabilitation services by diagnostic group and propensity to receive rehabilitation. 3. To evaluate the effect of these changes on survival and community discharge rate by comparing pre and post-PPS cohorts. Background: During the 1990's there was a rapid growth in the amount of rehabilitation therapy provided in NHs induced by discharge of patients from hospitals "sicker and quicker" and supported by changes in Medicare reimbursement policy. Patients who received rehabilitation services have been shown to have faired better in terms of discharge outcomes and survival than similar patients who did not receive such services. The new PPS for NHs reduces the financial incentives to NHs to provide rehabilitation services. This research will examine the magnitude of the change in rehabilitation services during the early post PPS period and the effect of the changes on important outcomes that have been reported in the past. Outcomes: Community discharge rates 3 months following the initial NH admission, operationalized by linking death records and longitudinally collected Minimum Data Set (MDS). Survival as operationalized by linking the MDS to the death records. Methods: Selection bias will be controlled using a previously developed propensity model employing 110 variables derived from the MDS. The propensity model will be developed and validated in a split pre-PPS sample and then applied to the patients in the post-PPS period to predict their propensity to receive these services, controlling for inter period selection differences. Susceptibility to long term NH residence and death will be directly controlled through 21 clinically chosen MDS variables. The primary analytic approaches will be case matching and multivariable modeling (logistic or proportional hazards). The success of the propensity modeling in controlling for selection bias will be confirmed using the case matching protocol. The study will provide valuable information concerning the results of a major public policy change in the funding of NH services.